A novel hybrid optimization model for evaluating and forecasting air quality grades.

Area of Science:

• Environmental Science
• Data Science
• Public Health

Background:

• Air pollution poses significant global risks to natural resources and public health.
• Accurate air pollution prediction is vital for implementing effective control strategies.
• Regional variations in primary pollutants present challenges for city-specific forecasting.

Purpose of the Study:

• To develop a novel method for accurate air pollution prediction tailored to regional variations.
• To categorize cities based on air quality index (AQI) values for targeted analysis.
• To enhance the accuracy of air quality grade prediction across multiple cities.

Main Methods:

• Integration of clustering algorithms to group cities by AQI.
• Application of functional principal component analysis (FPCA) to identify key pollution components within clusters.
• Utilization of an enhanced random forest algorithm for air quality grade prediction.

Main Results:

• The FP-RF model demonstrated high effectiveness in experimental evaluations.
• Achieved an average accuracy of 98.6% in forecasting six distinct pollution grades.
• Attained 96.04% accuracy in predicting air quality for 16 prefecture-level cities.
• Outperformed existing baseline models in prediction accuracy.

Conclusions:

• The proposed FP-RF method offers a robust and accurate approach to air pollution prediction.
• Clustering and FPCA effectively address regional variations in air pollution.
• The model provides a valuable tool for public health and environmental management.